robokudo.annotators.object_hypothesis_visualizer

Object hypothesis visualization for RoboKudo.

This module provides an annotator for visualizing object hypotheses in both 2D (image overlays) and 3D (point clouds) representations.

Classes

ObjectHypothesisVisualizer

Annotator for visualizing object hypotheses in the CAS.

Module Contents

class robokudo.annotators.object_hypothesis_visualizer.ObjectHypothesisVisualizer(name: str = 'ObjectHypothesisVisualizer', descriptor: ObjectHypothesisVisualizer = Descriptor())

Bases: robokudo.annotators.core.BaseAnnotator

Annotator for visualizing object hypotheses in the CAS.

This annotator creates visualizations of detected objects by:

  • Drawing bounding boxes and labels on the color image

  • Displaying associated point clouds in 3D

  • Optionally filtering objects based on query type

class Descriptor

Bases: robokudo.annotators.core.BaseAnnotator.Descriptor

Configuration descriptor for object hypothesis visualization.

class Parameters

Parameters for configuring visualization behavior.

query_aware: bool = True

If set to true, only visualize an Object that matches the ‘object.type’ from the Query.

visualize_full_cloud: bool = False

If set to true, the scene cloud will be shown and the individual objects will be colored.

parameters
_mesh_cache: typing_extensions.Dict[typing_extensions.Tuple[object, Ellipsis], open3d.geometry.TriangleMesh]
_iter_mesh_visuals(body: semantic_digital_twin.world_description.world_entity.Body) → typing_extensions.List[semantic_digital_twin.world_description.geometry.Mesh]

Return mesh-based visual shapes for a body.

_mesh_cache_key(shape: semantic_digital_twin.world_description.geometry.Mesh) → typing_extensions.Tuple[object, Ellipsis]

Create a stable cache key for mesh conversion results.

_mesh_shape_to_o3d(shape: semantic_digital_twin.world_description.geometry.Mesh) → open3d.geometry.TriangleMesh

Convert a mesh shape to Open3D, cache the result, and apply its local origin.

draw_text_middle(image: numpy.typing.NDArray[numpy.uint8], text: str, color: typing_extensions.Tuple[int, int, int] = (0, 0, 255), font: int = cv2.FONT_HERSHEY_SIMPLEX, font_scale: float = 1.0, thickness: int = 2)

Draw text in the middle of an image.

Parameters:

  • image – Image to draw on

  • text – Text to draw

  • color – BGR color tuple

  • font – OpenCV font type, defaults to cv2.FONT_HERSHEY_SIMPLEX

  • font_scale – Font scale factor, defaults to 1

  • thickness – Line thickness, defaults to 2

update() → py_trees.common.Status

Update the visualization with current object hypotheses.

Creates visualizations containing:

  • Color image with bounding boxes and labels for each object

  • 3D point clouds, pose and bounding box annotations for each object

  • Optional filtering based on query type if query_aware is True

Returns:

SUCCESS after creating visualizations